https://answers.netlify.com/t/limit-bandwidth-to-avoid-high-...
https://answers.netlify.com/t/limit-bandwidth-to-avoid-high-...
“The cool thing is that we also provide a load balancer, and if our system has detected that our main load balancer is currently being hit by a large DDoS attack and is slow or unresponsive, we’ll simply route around that on the DNS level. Since we cache content at our edge nodes around the world, end users also experience extremely fast page load times because of this.”
1. Comfortable paying >$100,000 to prevent their site going down for a single day, and
2. Doesn't pay a dime for their hosting service on a day-to-day basis
I don't understand why they won't just raise a 503 if the traffic exceeds the spend limit, or at the very least provide that as an option.
Autoscaling is a feature!
The issue at hand is that people put small websites on hosting providers designed for megacorporation wealth, like Netlify. I highly doubt the average blog needs more than a $10 VPS located in one single country without automatic fallback to another data centre. You can probably even go with a $5 VPS if you don't care about the first wave of HN front page bots not being able to reach your site.
I mean, yeah - but that shouldn't be the default and it shouldn't be something that you can't opt out of if it is, which is what sounds like happened with Netlify.
Same goes for most of the other pay-as-you-go providers that turn HN into billing support every now and then; very rarely do I see "we suddenly got a $20k bill" posts about services that these extreme availability products make sense for.
You wouldn't believe the amount of times I've said this and the response was "but it costs me nothing right now"...
Adding more regulation makes the system slower.
I'm fine with their pricing structure right now, since you have PLENTY of providers to choose from - people can easily vote with their wallets and there's no problem that needs solving.
However, the unexpected spikes are a problem, and providers seemingly don't provide any way to solve them because they make more money by not solving them. A regulation to require all providers of post-billed services to provide spending limits would make a lot of sense.
Of course, customers should also have the option to opt out of the limit or set a very high limit.
This should apply to any service that's billed by usage calculated afterwards, not just web hosting, and not just technology.
(But I also would like to see this feature)
I just had a look at the billing screen and I can see notifications but not budgeting options (on the Pro plan). Ideally I'd like to monitor/configure spend per deployment as some of our sites have big sales which can = big fees. Is this possible?
So it would be trivial for me to poll their budget API for an alert, and immediatly trigger a shutdown of my Cloudfront service. Why can't they do that for me?
Something based on this could be definitely better than nothing, but might also give false impression of safety.
[1] https://docs.aws.amazon.com/cost-management/latest/userguide...
I'm imagining an alert to the on-call team, and a soft shutdown until the on-call team can figure out the next step.
If it can save a few thousand dollars, it's worth it. Each business must make their own estimate of course.
Nope. Not in real time, they don't.
You can significantly overspend before the warning comes in.
I only realized this after Contabo contacted me and said the traffic is so high that other clients service is also degraded and they will have to take my VPS down if its much longer (which was understandable). Gladly the ddos stopped soon.
But never was there any talk about any cost, they were very supportive
But hey - just think about how much you saved on Netlify! Composable!
If their advertising is targeted to small businesses and individuals who could never afford this type of service, they could be guilty of false advertising, at least morally guilty. I haven't seen their marketing so I wouldn't want to say.
I've dealt with Netlify's support [1], and one of their CS heads was incredibly rude to me and blamed me for the problem they created.
https://docs.netlify.com/domains-https/custom-domains/config...